Next, I provided the exact same prompt to Gemini, which uses its new tool. The only difference is that I asked it to put a Google logo on the back of the woman's t-shirt. Here's what I got:
Okay, so it's not quite as good as OpenAI's version. The words aren't spelled as correctly, the reflection of the bridge isn't right, and even the t-shirt logo is misspelled. Still, it's pretty darn impressive, and we all have access to this model now.
